@Malco lol

Sup Guys, seen the new layout already?

Yes. :D

If you have feedback make sure to tag it design

@JonTheMon @JoeW @murgatroid99 I quickly updated the old one, may have missed things, let me know or edit further if you like.

Context?

(linked above)

Yeah, that looks like a good edit

Looks good to me.

@murgatroid99 Are you in here?

yes

I figured it was better to talk out a few points about the hand probability question here rather than in comments

ok

I agree that the answer to the new question isn't immediately obvious when looking at the old question, but it's certainly possible with some work. Even outside of that, isn't this the type of question we would close as being too specific and not useful to others?

I think it's actually more work to extend the answer to that old question to answer this new question than it was to answer the older question in the first place

And the close reason of being too specific, or "too localized" as it was called, was removed years ago. We don't do that anymore

It is, and I thought about doing it to show it was possible and decided it wasn't worth the effort on a question I thought should be closed

I did not realize we got rid of the "too localized" closure

I'm totally willing to let my close vote ride and see if other people agree. Outside of the tight focus and duplication, I didn't really have any other points.

I think that the analytical solution to this new question basically involves a very large, complex acyclic markov chain, where the transition probabilities can be calculated using the hypergeometric distribution described in the old answer

So, you can use the older answer, but doing the markov chain probability calculations is really the hard part there

Right. Does that belong on B&CG or Stas?

stats*

I mean, it's about MTG in a concrete sense, and I think the previous question gives a precedent that it is on topic

Anyway, it's more of a math question than statistics

Maybe I'm still stuck in the old mindset of "The answer of 95% isn't great for this site."

Knowing that "too localized" has been removed may have changed my mind on this question

Although anytime anyone has a different question about hand probabilities, we run the risk of having a new question that does some math and provides an answer.

Assuming there's something that alters those probabilities

I don't think there's anything wrong with that

Ok. I haven't really been an active user in the past few years and it seems like there's been a shift in a positive direction. I'm happy with that.

Vote retracted.

Thanks for clearing up some of my thoughts on this.

I started trying to write an analytical solution to that question but I don't know of a good way to calculate the probabilities of a markov chain with that many different states and transitions.

The basic problem is this: say you're considering the state as a triplet (C, S, H) representing the number of cards in the library, the number of serum powders in the library, and the number of cards you're drawing

C can range from 32 to 60, S ranges from 0 to 4 but some S values are incompatible with some C values, and H ranges from 1 to 7.

As for transitions, for any specific values (C, S, H) can transition to success, or (C-H, S-1, H), or (C-H, S-2, H), ..., or (C, S, H-1). And the probability of each transition depends on all 3 values using a hypergeometric distribution